Intercom Fin pricing explained: the $0.99 resolution, itemized

Intercom Fin pricing, as of July 2026, is $0.99 per resolution. A resolution counts when the customer confirms the answer helped, stops replying after Fin answers, or Fin completes a workflow. Intercom seats cost extra, from $29 per seat per month on the Essential plan with annual billing. Standalone Fin carries a monthly minimum.

Per resolution is the most honest meter in AI support, and Intercom publishes the rate, which is more than most vendors do. The questions that decide your invoice sit one layer down, starting with what Intercom counts as a resolution. This post answers them with the arithmetic visible.

How does the $0.99 per resolution model work?

You pay nothing for Fin conversations that do not resolve. When a conversation resolves, you pay $0.99, once, even if Fin answered several questions inside it. Intercom states both points on its pricing page as of July 2026. The seat plans are a separate charge on top, covered below.

What counts as a resolution?

Intercom bills per "outcome", its word for a resolution. Per the pricing page as of July 2026, an outcome is counted when:

• the customer confirms their issue is resolved, or
• the customer does not ask for more help after Fin responds, or
• Fin completes a workflow, which Intercom calls a Procedure, including handoffs.

Read the second one again. Silence counts. A customer who gets a wrong answer and closes the tab registers as a resolution and bills $0.99. Credit to Intercom for printing that on the pricing page. It still means your resolution count is an upper bound on your happy customers, and it means you should audit a sample of billed resolutions during your trial rather than trusting the topline rate.

The third one matters for stores. If you wire Fin into a Procedure that looks up an order, completing that Procedure counts as an outcome even when the conversation ends in a handoff to your team. Budget for that when you estimate your resolution share.

What do the seats cost on top?

Fin's meter runs on top of Intercom's seat plans. As of July 2026, seats start at $29 per seat per month on Essential with annual billing and run to $132 per seat per month on Expert. A two-person support team on Essential pays $58 per month before Fin resolves anything. The seat layer stays small next to the resolution layer at volume, but it is a floor, and it scales with your team rather than your tickets.

Can you run Fin without the Intercom helpdesk?

Yes. Fin works standalone on top of an existing helpdesk, with no seat costs, and a minimum monthly commitment applies. Intercom's pricing page gives 50 outcomes as the example minimum as of July 2026, which is about $50 per month at $0.99 each. Intercom does not publish a fixed minimum beyond that example, so confirm the exact commitment for your volume before signing.

What does Fin cost at 300, 1,000 and 2,500 conversations?

Assume a two-person team on Essential seats, $58 per month, and assume Fin resolves 60% of conversations. That 60% is an assumption. It sits in the range stores reach on routine questions, and it is the number to replace with your own trial data. The arithmetic:

• 300 conversations: 180 resolutions × $0.99 = $178.20. Plus seats: $58 + $178.20 = $236.20 per month.
• 1,000 conversations: 600 × $0.99 = $594. Plus seats: $652 per month.
• 2,500 conversations: 1,500 × $0.99 = $1,485. Plus seats: $1,543 per month.

Move the resolution share and the totals move nearly one for one, because the seats barely register at volume. At a 40% share, 1,000 conversations cost $396 in resolutions instead of $594. This is why the audit question above earns its time: every conversation billed as resolved that actually was not is $0.99 of leak.

Where per-resolution pricing is genuinely attractive

At low volume with high resolution quality, this model is hard to beat. A store with 150 conversations a month and a well-tuned Fin resolving 70% pays 105 × $0.99 = $103.95 plus seats, so around $162 all in. No tier jump, no shelf of unused capacity. You pay close to the value delivered, and a dollar for a resolved customer question is cheap against any human alternative.

Where the meter bites

The meter tracks your success. More orders produce more conversations, and better AI produces more resolutions, so the invoice grows on both axes at once. The sharp version is Q4. A store averaging 1,000 conversations that spikes to 2,500 in November pays $1,485 in resolutions that month instead of $594, at the same 60% share. The spike lands in the month your cash is already committed to inventory and ad spend, which is precisely when a 2.5x support bill hurts most.

A note for ecommerce stores

Fin is a general-purpose support agent, and Intercom serves every industry from SaaS to banking. It does not come wired to a store's data out of the box. Order status, returns and cancellations need Procedures connected to your Shopify or WooCommerce backend before Fin can act on them, which is real integration work rather than a toggle. Fin handles those flows once the plumbing exists. Price the build into your comparison, and remember that a completed Procedure bills as an outcome.

The flat-fee alternative

The other way to buy AI support is a flat monthly fee sized by volume. Nousu, our own product, works that way, so weigh this section accordingly. Three flat tiers as of July 2026: €29, €79 and €149 per month, sized for 150, 500 and 2,500 conversations. Going over a bundle never bills overage, so the busiest month costs the same as the slowest. The November spike above would leave a €149 invoice at €149. Nousu is also built for stores specifically: it reads live Shopify and WooCommerce order data and handles order status and returns without custom workflow builds.

Which model wins depends on your volume and your spike profile. Run both against your own numbers in the support cost calculator, and for the full map of AI support billing models, see our AI chatbot cost guide.

Fin is one of two big resolution meters worth understanding before you buy. Gorgias runs the other, with a per-ticket helpdesk meter underneath it and an unpublished AI rate on top. We took that one apart the same way in Gorgias pricing explained.